blob: 24e5ebcd0bcaefd7a8286e3f3d3445336174c669
1 | /* |
2 | * TwinVQ decoder |
3 | * Copyright (c) 2009 Vitor Sessak |
4 | * |
5 | * This file is part of FFmpeg. |
6 | * |
7 | * FFmpeg is free software; you can redistribute it and/or |
8 | * modify it under the terms of the GNU Lesser General Public |
9 | * License as published by the Free Software Foundation; either |
10 | * version 2.1 of the License, or (at your option) any later version. |
11 | * |
12 | * FFmpeg is distributed in the hope that it will be useful, |
13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
15 | * Lesser General Public License for more details. |
16 | * |
17 | * You should have received a copy of the GNU Lesser General Public |
18 | * License along with FFmpeg; if not, write to the Free Software |
19 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
20 | */ |
21 | |
22 | #ifndef AVCODEC_TWINVQ_H |
23 | #define AVCODEC_TWINVQ_H |
24 | |
25 | #include <math.h> |
26 | #include <stdint.h> |
27 | |
28 | #include "libavutil/common.h" |
29 | #include "libavutil/float_dsp.h" |
30 | #include "avcodec.h" |
31 | #include "fft.h" |
32 | #include "internal.h" |
33 | |
34 | enum TwinVQCodec { |
35 | TWINVQ_CODEC_VQF, |
36 | TWINVQ_CODEC_METASOUND, |
37 | }; |
38 | |
39 | enum TwinVQFrameType { |
40 | TWINVQ_FT_SHORT = 0, ///< Short frame (divided in n sub-blocks) |
41 | TWINVQ_FT_MEDIUM, ///< Medium frame (divided in m<n sub-blocks) |
42 | TWINVQ_FT_LONG, ///< Long frame (single sub-block + PPC) |
43 | TWINVQ_FT_PPC, ///< Periodic Peak Component (part of the long frame) |
44 | }; |
45 | |
46 | #define TWINVQ_PPC_SHAPE_CB_SIZE 64 |
47 | #define TWINVQ_PPC_SHAPE_LEN_MAX 60 |
48 | #define TWINVQ_SUB_AMP_MAX 4500.0 |
49 | #define TWINVQ_MULAW_MU 100.0 |
50 | #define TWINVQ_GAIN_BITS 8 |
51 | #define TWINVQ_AMP_MAX 13000.0 |
52 | #define TWINVQ_SUB_GAIN_BITS 5 |
53 | #define TWINVQ_WINDOW_TYPE_BITS 4 |
54 | #define TWINVQ_PGAIN_MU 200 |
55 | #define TWINVQ_LSP_COEFS_MAX 20 |
56 | #define TWINVQ_LSP_SPLIT_MAX 4 |
57 | #define TWINVQ_CHANNELS_MAX 2 |
58 | #define TWINVQ_SUBBLOCKS_MAX 16 |
59 | #define TWINVQ_BARK_N_COEF_MAX 4 |
60 | |
61 | #define TWINVQ_MAX_FRAMES_PER_PACKET 2 |
62 | |
63 | /** |
64 | * Parameters and tables that are different for each frame type |
65 | */ |
66 | struct TwinVQFrameMode { |
67 | uint8_t sub; ///< Number subblocks in each frame |
68 | const uint16_t *bark_tab; |
69 | |
70 | /** number of distinct bark scale envelope values */ |
71 | uint8_t bark_env_size; |
72 | |
73 | const int16_t *bark_cb; ///< codebook for the bark scale envelope (BSE) |
74 | uint8_t bark_n_coef;///< number of BSE CB coefficients to read |
75 | uint8_t bark_n_bit; ///< number of bits of the BSE coefs |
76 | |
77 | //@{ |
78 | /** main codebooks for spectrum data */ |
79 | const int16_t *cb0; |
80 | const int16_t *cb1; |
81 | //@} |
82 | |
83 | uint8_t cb_len_read; ///< number of spectrum coefficients to read |
84 | }; |
85 | |
86 | typedef struct TwinVQFrameData { |
87 | int window_type; |
88 | enum TwinVQFrameType ftype; |
89 | |
90 | uint8_t main_coeffs[1024]; |
91 | uint8_t ppc_coeffs[TWINVQ_PPC_SHAPE_LEN_MAX]; |
92 | |
93 | uint8_t gain_bits[TWINVQ_CHANNELS_MAX]; |
94 | uint8_t sub_gain_bits[TWINVQ_CHANNELS_MAX * TWINVQ_SUBBLOCKS_MAX]; |
95 | |
96 | uint8_t bark1[TWINVQ_CHANNELS_MAX][TWINVQ_SUBBLOCKS_MAX][TWINVQ_BARK_N_COEF_MAX]; |
97 | uint8_t bark_use_hist[TWINVQ_CHANNELS_MAX][TWINVQ_SUBBLOCKS_MAX]; |
98 | |
99 | uint8_t lpc_idx1[TWINVQ_CHANNELS_MAX]; |
100 | uint8_t lpc_idx2[TWINVQ_CHANNELS_MAX][TWINVQ_LSP_SPLIT_MAX]; |
101 | uint8_t lpc_hist_idx[TWINVQ_CHANNELS_MAX]; |
102 | |
103 | int p_coef[TWINVQ_CHANNELS_MAX]; |
104 | int g_coef[TWINVQ_CHANNELS_MAX]; |
105 | } TwinVQFrameData; |
106 | |
107 | /** |
108 | * Parameters and tables that are different for every combination of |
109 | * bitrate/sample rate |
110 | */ |
111 | typedef struct TwinVQModeTab { |
112 | struct TwinVQFrameMode fmode[3]; ///< frame type-dependent parameters |
113 | |
114 | uint16_t size; ///< frame size in samples |
115 | uint8_t n_lsp; ///< number of lsp coefficients |
116 | const float *lspcodebook; |
117 | |
118 | /* number of bits of the different LSP CB coefficients */ |
119 | uint8_t lsp_bit0; |
120 | uint8_t lsp_bit1; |
121 | uint8_t lsp_bit2; |
122 | |
123 | uint8_t lsp_split; ///< number of CB entries for the LSP decoding |
124 | const int16_t *ppc_shape_cb; ///< PPC shape CB |
125 | |
126 | /** number of the bits for the PPC period value */ |
127 | uint8_t ppc_period_bit; |
128 | |
129 | uint8_t ppc_shape_bit; ///< number of bits of the PPC shape CB coeffs |
130 | uint8_t ppc_shape_len; ///< size of PPC shape CB |
131 | uint8_t pgain_bit; ///< bits for PPC gain |
132 | |
133 | /** constant for peak period to peak width conversion */ |
134 | uint16_t peak_per2wid; |
135 | } TwinVQModeTab; |
136 | |
137 | typedef struct TwinVQContext { |
138 | AVCodecContext *avctx; |
139 | AVFloatDSPContext *fdsp; |
140 | FFTContext mdct_ctx[3]; |
141 | |
142 | const TwinVQModeTab *mtab; |
143 | |
144 | int is_6kbps; |
145 | |
146 | // history |
147 | float lsp_hist[2][20]; ///< LSP coefficients of the last frame |
148 | float bark_hist[3][2][40]; ///< BSE coefficients of last frame |
149 | |
150 | // bitstream parameters |
151 | int16_t permut[4][4096]; |
152 | uint8_t length[4][2]; ///< main codebook stride |
153 | uint8_t length_change[4]; |
154 | uint8_t bits_main_spec[2][4][2]; ///< bits for the main codebook |
155 | int bits_main_spec_change[4]; |
156 | int n_div[4]; |
157 | |
158 | float *spectrum; |
159 | float *curr_frame; ///< non-interleaved output |
160 | float *prev_frame; ///< non-interleaved previous frame |
161 | int last_block_pos[2]; |
162 | int discarded_packets; |
163 | |
164 | float *cos_tabs[3]; |
165 | |
166 | // scratch buffers |
167 | float *tmp_buf; |
168 | |
169 | int frame_size, frames_per_packet, cur_frame; |
170 | TwinVQFrameData bits[TWINVQ_MAX_FRAMES_PER_PACKET]; |
171 | |
172 | enum TwinVQCodec codec; |
173 | |
174 | int (*read_bitstream)(AVCodecContext *avctx, struct TwinVQContext *tctx, |
175 | const uint8_t *buf, int buf_size); |
176 | void (*dec_bark_env)(struct TwinVQContext *tctx, const uint8_t *in, |
177 | int use_hist, int ch, float *out, float gain, |
178 | enum TwinVQFrameType ftype); |
179 | void (*decode_ppc)(struct TwinVQContext *tctx, int period_coef, int g_coef, |
180 | const float *shape, float *speech); |
181 | } TwinVQContext; |
182 | |
183 | extern const enum TwinVQFrameType ff_twinvq_wtype_to_ftype_table[]; |
184 | |
185 | /** @note not speed critical, hence not optimized */ |
186 | static inline void twinvq_memset_float(float *buf, float val, int size) |
187 | { |
188 | while (size--) |
189 | *buf++ = val; |
190 | } |
191 | |
192 | static inline float twinvq_mulawinv(float y, float clip, float mu) |
193 | { |
194 | y = av_clipf(y / clip, -1, 1); |
195 | return clip * FFSIGN(y) * (exp(log(1 + mu) * fabs(y)) - 1) / mu; |
196 | } |
197 | |
198 | int ff_twinvq_decode_frame(AVCodecContext *avctx, void *data, |
199 | int *got_frame_ptr, AVPacket *avpkt); |
200 | int ff_twinvq_decode_close(AVCodecContext *avctx); |
201 | int ff_twinvq_decode_init(AVCodecContext *avctx); |
202 | |
203 | #endif /* AVCODEC_TWINVQ_H */ |
204 |